Evaluating Magma
Chris Muller
asqueaker at gmail.com
Sun Oct 7 22:59:42 UTC 2007
Thanks for your cents Sebastian.
I must agree with:
> newcomers will allways found the default behavior of
> #refreshPersistentObjectsEvenWhenChangedOnlyByMe in false
> as a strange thing.
because this question is asked a lot. Like you, everyone seems to be
surprised by this.
But, notwithstanding the above the nature of some applications will
insist the users work not be disturbed upon crossing a transaction
boundary or commit failure. The default to this preference is false
because it is more conservative with the users work and also performs
better.
> Heuristically is preferable that the default where true
Why?
> Questions (I didn't saw in FAQ):
>
> . I know I'll probably do not implement (nor want to) it but I need to
> know if it's planned to support pessimistic locking somtime.
I have no plans to do it. But could this be handled at the
application-level anyway?
> . there is an idea of how many concurrent clients (sessions) can be
> served by a squeak image? or any other metrics relevant to measure load
> capacity/scalability? OK I've re-read Magma limitations, there is any news
> on this?
Since there are so many variables in this question, check out the
Magma performance benchmarks page which may help you answer for your
intended use..
http://wiki.squeak.org/squeak/5606
> . there is an idea of how much is Magma resistant to power faliures?
> (sata drives with software raid)
It is very robust in this area. Partial writes are detected and
recovered automatically.
- Chris
More information about the Magma
mailing list